GS-PRM: Gaussian Visual Roadmap for Image-Goal Navigation
We introduce \method, a Gaussian visual roadmap whose milestones, local connections, and visual references are constructed directly through 3DGS rendering without deriving an occupancy grid or signed distance field.
ProofPaper ↗
Key points
- Image-goal navigation requires a global planner that connects visually specified endpoints while providing observations that a local policy can follow.
- Three-dimensional Gaussian Splatting (3DGS) provides map-registered RGB, opacity, and depth renderings, but does not explicitly encode navigable topology.
- We define render-based collision-free (RC-free) configurations and paths.
- To validate the performance of GS-PRM, we conduct experiments in both simulated and real-world environments, demonstrating long-horizon navigation using goal-conditioned policies guided by GS-PRM.
Sources (1)
- [1]GS-PRM: Gaussian Visual Roadmap for Image-Goal NavigationarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 14, 03:58 AM
We introduce \method, a Gaussian visual roadmap whose milestones, local connections, and visual references are constructed directly through 3DGS rendering without deriving an occupancy grid or signed distance field.
Image-goal navigation requires a global planner that connects visually specified endpoints while providing observations that a local policy can follow.
Extractive summary: sentences quoted from the sources.